Here is my fix.
I assumed that a evaporating handle allows the monitor thread to resign. The consequence should be that it gets recreated with a reopened device handle, but I am not sure about this.
The patch solves the 100% cpu load polling, but as already said I am not sure if you loose the Monitor this way further on.
Reading your patch suggestion I'd say that not checking poll on it's return value is to consider a bug.
Here is my fix.
I assumed that a evaporating handle allows the monitor thread to resign. The consequence should be that it gets recreated with a reopened device handle, but I am not sure about this.
The patch solves the 100% cpu load polling, but as already said I am not sure if you loose the Monitor this way further on.
Reading your patch suggestion I'd say that not checking poll on it's return value is to consider a bug.
Best
Helmut